OXYGENHUB
◆ paths-for-resolutions-and-amend oxygen / crates / oxy-core / src / repo / gc.rs
9.8 KiB · Rust · last changed Readers never wait: one head per handle, retry on a miss, retire before deleting 17 days ago source blame history raw
1 
//! Garbage collection (spec §21): delete objects
2 
//! nothing reachable refers to.
3 
//!
4 
//! Roots: the current view, the views of every operation newer than
5 
//! `core.op-retention` (their working copies only within the shorter
6 
//! `core.snapshot-retention`), reference attestations, and the policy; receipts
7 
//! live as long as the changes of their state. Operation
8 
//! objects are always kept so the log stays walkable, but an operation older
9 
//! than the retention window no longer keeps its states alive, so undoing
10 
//! that far back may fail once its objects are gone.
11
12 
use super::*;
13 
use crate::store::{PackReport, PruneReport, RetiredReport};
14 
use std::collections::HashSet;
15
16 
#[derive(Debug, Clone, Default, PartialEq, Eq)]
17 
pub struct GcReport {
18 
    pub packed: PackReport,
19 
    pub pruned: PruneReport,
20 
    pub reachable: usize,
21 
    /// Packs folded into the one a full repack writes.
22 
    pub merged_packs: usize,
23 
    /// Packs and loose objects earlier repacks retired, deleted now that
24 
    /// [`crate::store::RETIRED_FOR`] has passed (oxygenhub#249).
25 
    pub retired: RetiredReport,
26 
}
27
28 
impl Repo {
29 
    /// Delete what earlier repacks retired more than
30 
    /// [`crate::store::RETIRED_FOR`] ago, prune unreachable objects older
31 
    /// than the grace period (`now` ignores it, but not the retired
32 
    /// files' hour), then repack everything into one pack (spec §13.4).
33 
    /// The packs and loose objects this repack replaces are retired in
34 
    /// turn, for the next collection.
35 
    pub fn gc(&self, now: bool) -> Result<GcReport> {
36 
        self.gc_with(now, &mut |_| {})
37 
    }
38
39 
    /// [`Repo::gc`], reporting each phase and the packing's progress
40 
    /// (oxygen#66), for a spinner or a log.
41 
    pub fn gc_with(&self, now: bool, progress: crate::store::Progress) -> Result<GcReport> {
42 
        self.gc_keeping(now, &[], progress)
43 
    }
44
45 
    /// [`Repo::gc_with`], keeping `extra` blobs and their chunks as well:
46 
    /// objects a host stored for itself and names from outside the
47 
    /// repository's history, such as a forge's job artifacts
48 
    /// (oxygenhub#126).
49 
    pub fn gc_keeping(
50 
        &self,
51 
        now: bool,
52 
        extra: &[ObjectRef],
53 
        progress: crate::store::Progress,
54 
    ) -> Result<GcReport> {
55 
        // Retired files first: the readers that used them have had their
56 
        // hour, and what replaced them is in the packs read from here on.
57 
        let retired = self
58 
            .store
59 
            .expire_retired(crate::store::RETIRED_FOR, progress)?;
60 
        progress("finding what is reachable");
61 
        let mut keep = self.reachable()?;
62 
        for r in extra {
63 
            if !self.store.has(r) || !keep.insert(*r) {
64 
                continue;
65 
            }
66 
            if let Ok(body) = self.store.get_typed(r, ObjectType::Blob) {
67 
                if let Ok(blob) = Blob::decode(&body) {
68 
                    keep.extend(blob.chunks);
69 
                }
70 
            }
71 
        }
72 
        let grace = if now {
73 
            std::time::Duration::ZERO
74 
        } else {
75 
            std::time::Duration::from_secs(self.config.gc_grace_days * 86_400)
76 
        };
77 
        let pruned = self.store.prune_with(&keep, grace, progress)?;
78 
        // Then everything left in one pack, compressed harder than a
79 
        // commit's (spec §13.4), as `git gc` repacks.
80 
        let (packed, merged_packs) = self
81 
            .store
82 
            .repack_all_with(crate::store::REPACK_ZSTD_LEVEL, progress)?;
83 
        Ok(GcReport {
84 
            packed,
85 
            pruned,
86 
            reachable: keep.len(),
87 
            merged_packs,
88 
            retired,
89 
        })
90 
    }
91
92 
    /// Every object that must survive garbage collection.
93 
    pub fn reachable(&self) -> Result<HashSet<ObjectRef>> {
94 
        let mut keep: HashSet<ObjectRef> = HashSet::new();
95 
        let mut states: Vec<ObjectRef> = Vec::new();
96 
        let mut changes: Vec<ChangeHash> = Vec::new();
97 
        let mut attests: Vec<ObjectRef> = Vec::new();
98 
        let mut policies: Vec<ObjectRef> = Vec::new();
99
100 
        // Operations: always kept; views within retention are roots. Past
101 
        // working copies only within the shorter `core.snapshot-retention`
102 
        // (oxygen#98): a file only ever snapshotted, such as a build output
103 
        // a status took in, goes once they expire.
104 
        let now_ms = Hlc::now().wall_ms;
105 
        let cutoff = now_ms.saturating_sub(self.config.op_retention_days * 86_400_000);
106 
        let wc_cutoff = now_ms
107 
            .saturating_sub(self.config.snapshot_retention_days * 86_400_000)
108 
            .max(cutoff);
109 
        for (i, (r, op)) in self.op_log()?.into_iter().enumerate() {
110 
            keep.insert(r);
111 
            if i == 0 || op.hlc.wall_ms >= cutoff {
112 
                states.extend(op.view.bookmarks.values().copied());
113 
                states.extend(op.view.remotes.values().flat_map(|b| b.values().copied()));
114 
                policies.extend(op.view.policy);
115 
            }
116 
            if i == 0 || (self.config.snapshot_retention_days > 0 && op.hlc.wall_ms >= wc_cutoff) {
117 
                changes.extend(op.view.workspaces.values().copied());
118 
            }
119 
        }
120 
        // Checkpoints: the object, its payload blob, and the blob's chunks.
121 
        for r in self.checkpoint_refs()? {
122 
            if let Ok(bytes) = self.store.get_typed(&r, ObjectType::Checkpoint) {
123 
                keep.insert(r);
124 
                let cp = crate::objects::Checkpoint::decode(&bytes)?;
125 
                keep.insert(cp.payload);
126 
                if let Ok(b) = self.store.get_typed(&cp.payload, ObjectType::Blob) {
127 
                    keep.extend(Blob::decode(&b)?.chunks);
128 
                }
129 
                // A version 2 payload's graph blobs and their chunks.
130 
                for g in self.checkpoint_graph_blobs(&cp).unwrap_or_default() {
131 
                    keep.insert(g);
132 
                    if let Ok(b) = self.store.get_typed(&g, ObjectType::Blob) {
133 
                        keep.extend(Blob::decode(&b)?.chunks);
134 
                    }
135 
                }
136 
            }
137 
        }
138 
        // Tags: their states and attestations.
139 
        let view_now = self.view()?;
140 
        states.extend(view_now.tags.values().copied());
141 
        for name in view_now.tags.keys() {
142 
            if let Some((r, _)) = self.tag_attest(name)? {
143 
                attests.push(r);
144 
            }
145 
        }
146 
        // Reference attestations.
147 
        for name in self.view()?.bookmarks.keys() {
148 
            if let Some((r, _)) = self.latest_ref_attest(name)? {
149 
                // With a host's acceptance of it (spec §14.7).
150 
                if let Some(acc) = self.acceptance_of(&r)? {
151 
                    attests.push(acc.attest.hash());
152 
                }
153 
                attests.push(r);
154 
            }
155 
        }
156 
        // A historical view may name objects pruned under an earlier, shorter
157 
        // retention; there is nothing to keep for those.
158 
        for st in states {
159 
            if keep.insert(st) {
160 
                match self.store.get_state(&st) {
161 
                    Ok(s) => changes.extend(s.heads),
162 
                    Err(StoreError::NotFound(_)) => continue,
163 
                    Err(e) => return Err(e.into()),
164 
                }
165 
            }
166 
        }
167 
        for p in policies {
168 
            if keep.insert(p) {
169 
                let bytes = match self.store.get_typed(&p, ObjectType::Policy) {
170 
                    Ok(b) => b,
171 
                    Err(StoreError::NotFound(_)) => continue,
172 
                    Err(e) => return Err(e.into()),
173 
                };
174 
                let policy = crate::policy::Policy::decode(&bytes)?;
175 
                for id in &policy.identities {
176 
                    keep.insert(id.hash());
177 
                }
178 
                for sig in self.store.signatures(&p)? {
179 
                    keep.insert(sig.key);
180 
                }
181 
            }
182 
        }
183 
        // Changes: their closure, identities, blobs, chunks, seals.
184 
        let mut stack = changes;
185 
        while let Some(h) = stack.pop() {
186 
            if !keep.insert(h) {
187 
                continue;
188 
            }
189 
            // Beyond the shallow boundary, or pruned under an earlier
190 
            // retention, there is nothing to keep.
191 
            let c = match self.store.get_change_parts(&h) {
192 
                Ok((c, texts)) => {
193 
                    keep.extend(texts);
194 
                    c
195 
                }
196 
                Err(StoreError::NotFound(_)) => continue,
197 
                Err(e) => return Err(e.into()),
198 
            };
199 
            stack.extend(c.deps.iter().copied());
200 
            if !c.author.is_zero() {
201 
                keep.insert(c.author);
202 
            }
203 
            for sig in self.store.signatures(&h)? {
204 
                self.keep_signer(&mut keep, sig.key)?;
205 
            }
206 
            for e in &c.edits {
207 
                if let Edit::Replace { old, new, .. } = e {
208 
                    for b in [old, new].into_iter().flatten() {
209 
                        if keep.insert(*b) {
210 
                            if let Ok(bytes) = self.store.get_typed(b, ObjectType::Blob) {
211 
                                keep.extend(Blob::decode(&bytes)?.chunks);
212 
                            }
213 
                        }
214 
                    }
215 
                }
216 
            }
217 
            attests.extend(self.store.seals_for(&h)?);
218 
        }
219 
        self.reachable_receipts(&mut keep)?;
220 
        self.reachable_approvals(&mut keep)?;
221 
        for a in attests {
222 
            if keep.insert(a) {
223 
                for sig in self.store.signatures(&a)? {
224 
                    self.keep_signer(&mut keep, sig.key)?;
225 
                }
226 
            }
227 
        }
228 
        Ok(keep)
229 
    }
230
231 
    /// Keep a signer's identity and, for a short-lived key, its
232 
    /// certificate and the certificate's issuer (spec §14.7).
233 
    pub(crate) fn keep_signer(&self, keep: &mut HashSet<ObjectRef>, key: ObjectRef) -> Result<()> {
234 
        keep.insert(key);
235 
        if let Some(c) = self.store.certificate_of(&key)? {
236 
            keep.insert(c);
237 
            if let Ok(cert) = self.store.get_certificate(&c) {
238 
                keep.insert(cert.cert.issuer);
239 
            }
240 
        }
241 
        Ok(())
242 
    }
243 
}